Topic: Suggestions for upgrading rig

[F.O.Bean]

LB>M10

[Brian E.]

yeah, I truly believe that going 24-bit is the best thing you can do, for a start.  It gives you a lot more headroom for boosting and editing later, so that you can run a little low and still get a good recording.

[F.O.Bean]

yeah, I truly believe that going 24-bit is the best thing you can do, for a start.  It gives you a lot more headroom for boosting and editing later, so that you can run a little low and still get a good recording.

I agree

[preert]

Today, I use: Rode NT4 -> Denecke AD20 -> iriver 140 (battery&memory card modded).I'm happy with the current rig, but I kinda like buing new stuff. What would you suggest for my improving the above rig?^^^ never heard that around here before.lol

[OOK]

LB>M10

+



=
 

[F.O.Bean]

LB>M10

+



=
 

Once again, agreed

[matsa]

Yep, message received. That is a killer combo.
But... I have located a couple of used fr-2le in my area, and I think I'll get one.
This way I also go from two boxes to one.
At a later point I'll look into the stereo mics from LineAudio design (ST6Li, http://www.lineaudio.se/linemic.htm ) as well as the BP4025.

Thanks again!

[acidjack]

Interesting mics... and very inexpensive at ~$160USD each..

[SmokinJoe]

If you want low self noise out of a mic, Large Diaphragm Condensers are generally quieter than Small Diaphragm condensers, per the laws of physics.  I'm not saying all LD's are quieter than all SD's without exception, but within a similar price range, it is frequently true.  The flip side is that the mic tends to be bigger and heavier.  I ran an LSD2 > PS2 / AD20 > H120 a few times, and it sounded pretty nice.  Yes, that's a phantom powered mic, and it's big and heavy.  The AD-20 is generally regarded as a pretty quiet pre, and very easy on batteries.

You mentioned the AT4025... I've never heard one, but if it's similar to the AT825, that was a decent mic, but I wouldn't rank it as a huge upgrade over the NT4.

Most of the improvements in a rig are made on that analog side of the chain... mics + preamp... right up to the A/D converter, and sometimes including the A/D.  Having a 24bit or 16bit "bit bucket" is much more subtle than most people think.  Breaking down the sample into finer resolution isn't nearly as important as having a good signal to sample. 

The other big thing is that mics and preamps have "flavor" which is difficult to describe. Some combos go together great (like peanut butter and jelly), and some are terrible together (like oil and water).  Predicting the good combos versus the bad combos isn't always easy.  If you spend a lot of time on archive.org, you will find certain combos are used a lot.... So what does that tell you?  People tried it and they liked it, so they kept it and ran it for years.  Pick some other combo at random, and you won't find many sources... maybe no one has tried it, or maybe they did try it, it sucked, and they bought something else after a couple of shows... Darwinian selection.  That's the advantage to finding a "tried and true" combo.  If you let people like us throw out random suggestions and put those pieces together, it's anybody's guess how they might sound, or if you will find that pleasing.  But we do get to learn if a combo works or not, and you get to pay for it.

[acidjack]

^^ QFT

[F.O.Bean]

^^ QFT

[Gutbucket]

Most of the improvements in a rig are made on that analog side of the chain... mics + preamp... right up to the A/D converter, and sometimes including the A/D.  Having a 24bit or 16bit "bit bucket" is much more subtle than most people think.  Breaking down the sample into finer resolution isn't nearly as important as having a good signal to sample.

(my bold above)  Just picking nits in a spam bumped thread..

I agree in a practical sense with everything mentioned in Joe's post, but just want to clear up a common technical misconception in regards to the signifigance of bit depth and sampling rate.  I think the phrase 'finer resolution' often misleads people who do not have a strong understanding of how PCM sampling works.  As long as the bit depth used is sufficient to cover the dynamic range of the analog signal being recorded, increasing the bit depth from 16 to 24 bits does not mean storing smaller, finer, or more accurate incremental values of loudness within that range.  Instead, increasing the bit depth increases the total number of recordable values, which then cover a proportionally larger range of signal levels.  A higher bit rate is capable of recording a wider overall range of signal levels.  That is advantageous to us as recordists because we do not know beforehand what the absolute signal levels will be, so extra unused room at the top and bottom of the recorded range can be especially useful, even if the extent of that increased range is far greater than need to reproduce the source signal with the same acuracy.

Similarly, a higher sampling rate does not record more information than a lower sampling rate does within the frequency range covered by the lower sampling rate, it extends the recorded frequency range to higher frequencies.  Higher bit rates extend the overall range of dynamic levels.

[acidjack]

^^^ In layman's terms, does that not (possibly) mean:

"If recording a PA system with limited dynamic range, a lower bit depth (say, 16bit) and sampling rate (say, 44.1) is likely sufficient to capture the full existing frequency range"?

I'm not arguing the merits of what anyone does - I record at 24/48 myself "just to be sure" - but scientifically, isn't that the case?  Unless of course PA systems have much higher quality than I think they do....

[aaronji]

"If recording a PA system with limited dynamic range, a lower bit depth (say, 16bit) and sampling rate (say, 44.1) is likely sufficient to capture the full existing frequency range"?

I think this conflates dynamic range and frequency range a little (at least as written).  For a source with limited dynamic range, a lower bit depth should be sufficient to capture it (16 bits is ~ 96 dB dynamic range).  The bit depth doesn't really have anything to do with the frequency range, though.  That's where the sampling rate comes in.  Basically, you can completely reconstruct the waveform for frequencies up to the Nyquist frequency (one half of the sampling frequency).  So at 44.1, you are good up to a frequency of 22.05 kHz.  At 48, you are good up to 24 kHz.  Human hearing is generally quoted as going up to 20 kHz, so you could argue that any sampling frequency past 44.1 or 48 is unlikely to yield any audible benefit...

So, really, assuming I actually understand this stuff, the above quote should be split into two parts:
"If recording a PA system with limited dynamic range, a lower bit depth (say, 16bit) is likely sufficient to capture the full existing dynamic range" and
"If recording a PA system with limited frequency range, a lower sampling rate (say, 44.1) is likely sufficient to capture the full existing frequency range"

[Gutbucket]

My personal/practical take away is that the 16/44.1 PCM in itself is entirely sufficient as a reproduction format for any music source, including highly dynamic material with a full frequency response.  Yet I feel there are real advantages for recording and processing at higher rates and wider  bandwidths, within practical limits of diminishing returns.  I've settled on recording raw files at 24/48 and the software I use for processing works at a further increased bit depth and somtimes oversamples internally for its calculations.  Yet once all that is complete and the recording is optimized for playback, I don't feel the difference between 16/44.1 and higher rates is significant enough for me to be very concerned about.

In layman's terms, it's about sufficient working space and wiggle room. 

Care for a few tortured analogies? My finished acoustic guitar sounds great and fits in a snug case which is the perfect size to hold it, a bigger case wouldn't make it sound any better, but there is no way I could build an equivalent guitar inside the case, I'd need much more workspace to do it right.  Yet once completed, the small case is perfect and more efficient than storing the guitar in a dedicated room.

There is a precise amount of gas required to get your car from point A to point B, however the practical question in the real world is rather "am I sure I have enough gas to make it?" How close are you willing to cut it before running out of gas?

I brought this up only to correct what I think is a common misconception about a clear black and white technical aspect of how the format works.  In practical 'real world' implementations, subjective perceptions and all kinds of confusing complications arise. A few- Yes a 44.1 sampling rate can theoretically record and reproduce signal frequencies of up to 22 kHz.  Yet to do so the frequencies above 22kHz must be completely filtered out, and real world practical filters which don’t sound bad require some range over which to work, so rolling off starting at 18 or 20kHz might be a better engineering choice.  Theoretically a 16bit PCM file has a dynamic capability of up to 96dB, but how many ’24 bit’ recorders that we use have an actual dynamic range capability of much more or even that much?  Different implementations, chips, filters and circuits both measure different and sound subjectively different to people, but that’s the down and dirty practical engineering and subjective side of things..  which people argue endlessly.